Fatique failure due to vibration

[Rattletrap]

I am trying to solve a problem regarding a bracket that was built and used on a truck. The bracket lasted only 2 weeks. The bracket is made of 316 stainless steel. I ran a vibration analysis on it and found that its natural frequency is about 50 Hz in the 6 scenarios shown by the software.

I would like to know how to use this information. I thickened the bracket and the frequency inreased to about 100 and 120 Hz. Is this a good thing? Can anyone explain what is happening here and how to solve the problem. What should the frequency be? Which is best a low or high frequency. I understand that the natural frequency of the bracket should not match the forced vibration from the truck but is it as simple as that?

Thanks in advance for your help

 

[desertfox]

Hi Rattletrap

Can you provide more information about the failure ie a picture of the bracket, where it fits on the truck, a picture of the failed bracket and we maybe able to advise further.
Your right though you need to have the natural frequency of the bracket as far away as possible from the forced vibration from the truck.

desertfox

 

[ivymike]

is the bracket attached to the engine, or is it elsewhere on the truck?

 

[GregLocock]

I'd add that 316 is a fairly bad choice most of the time.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[IRstuff]

some engines just seem to be overly torquey for their own good. I used to have a relatively wimpy 2.2L engine that routinely broke its front engine mount. All three were even OEM mounts bought directly through the dealer. Never could understand how the motor managed to break the mount, but it did, right in the middle of about 2 inches of steel folded into an "L" shape.

TTFN

FAQ731-376

 

[Rattletrap]

It proberbly fits on the body on the front since it holds a spotlight.

The picture I am adding is of the bracket after I made changes to it already. But I am not convinced that this will guarantee that it would not fail again.

What material will be better than 316 Stainless steel?

Thanks again for the help

 

[desertfox]

Hi Rattletrap

The points at which it failed, was there a nut and bolt at that position? it seems to me that its possible the nut and bolt as come loose and allows the spotlight to vibrate.
What changes have you made to the bracket.

desertfox

 

[Tmoose]

Does your model include the body attachment point, and the mass of the spotlight with some representation of the spotlight's grip on the bracket?

 

[Rattletrap]

The bolts are held in with loctite. And yes it failed close to the bolt clamping area where the light is mounted see the picture above. The attachement point and the grip on the bracket is all visible on the picture above. The mass of the Spot light is proberbly aroound 3 kg or so. The sheet thickness was 4 mm before I changed it to 6 mm now and I see I get a higher frequency up to 200 Hz with the new design after I added a brace and angle piece on the base of the bracket.
The question still remains, what is best a higher frequence or a lower frequency. I wonder if my conclusion is correct by saying that a higher frequency mean that the actual movement of the part will also be less than if the frequence was low and therefor decreases fatique because the stress is less?

Thanks again.

 

[desertfox]

hi Rattletrap

I think you might be better increasing the braket depth ie increase the material thickness from the edge of the circular hole pattern to the centre of the holes.
Even loctite might not stop the fixings working loose.

desertfox

 

[1gibson]

The circle of holes is for aiming the spotlights up/down, yes?

How likely is it that you will ever aim them straight up or straight down? Eliminate 3-4 holes at the very top and very bottom of the bracket, where the failure was. Reduce the stress concentration there.

Does the bracket get particularly hot?

 

[GregLocock]

316 generally has poor fatigue properties compared with non SS steels.

Yeah make it thicker etc, I'd put a doubler in around those holes.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Tmoose]

If my model is good it will show me where to add or remove material to produce lowest stress for various loadings and vibration modes.

What was the modeshape of the 50 Hz resonance?
Like "a" here?

http://www.lhup.edu/~dsimanek/scenario/labman2/tuneFORK.gif

If so, I think the spotlight shell may not being represented accurately.

The boundary conditions of mounting must reasonably approximate the real condition to get usefully accurate natural frequencies. A free-free model of the bracket without the "probably 3kg" spotlight would not appoximate the real mounted system in any useful way. A realistically mounted (secured to a plane by patches around the screws) bracket but without the mass and stiffness influences of the spotlight would have higher than realistic resonant frequencies, and some unrealistic modes, like the one in th elink above.

If I'm operating very near a resonance then with minimal damping the amplitude is theoretically infinite. REal life amplification of 10X or more is EZ. Not many designs have a 10X margin in safety factor, or materials' endurance limit.
The rapid failure suggests resonance may be involved.

Vibration measurements of the system in operation would identify important frequencies and save a lot of time.

 

[rob768]

If you get a low frequency, you have a weak construction (or a large active mass). This may be in one direction only, or in general. A weak construction usually relates to high deflections and if your structure has an area of stress concentrations (for instance the sharp corners in your construction) you will end up with high stress. With a progressive change in stiffness, this effect is smaller.
A high frequency is either a light or a stiff construction.
Either can be a problem if a potentially bothersome natural frequency is close to an excitation source, for instance combustion frequency of the engine.

 

[ColonelSanders83]

If I wanted to fail something in fatuige i would perferate it as you have shown.

If I was asked to remake the part I would change the design so that the plate was not perferated (spot weld a perferated plat to the back, ect).

Natural frequency shifts can only help you if you were having a resonance issue in the first failure. If you have very high vibration levels, shifts in natural frequency will be unlikley to up the lift of the part.

Not a ding on you, but every once in a while you come accross something that is a poor design at which point the best bet is to start over (in this case without perforations)

Also the natural frequency of just the bracket is rather useless, you need to know the natural frequence of the bracket when installed (the whole assembly)

Just my two cents worth.

A question properly stated is a problem half solved.

Always remember, free advice is worth exactly what you pay for it!

http://www.ap-dynamics.ab.ca/

 

[hacksaw]

Are you certain that this is a really fatigue failure? I am having difficulty visualizing how the loads are applied to the bracket.

 

[izax1]

To your original question: Low frequency or high frequency?

Since you are considering fatigue failure, just remember that lower frequencies will give generally higher deformation and stresses but less fatigue cycles, while the opposite is the case for hihger freuencies. Reducing stresses must be considered against increased No of Cycles. It really all depends on your input. In what frequency band is the high input level?

Generally, I would say that higher frequency is better, but that is not the complete answer. You also need to consider input levels, material fatigue properties, required fatigue life, stress concentrations etc.

 

[DarrellThomasNV]

That is a lot of holes around the edges. What are they for? is that where a locking pin goes to lock the spotlight in place?

It looks like your crack propagated from hole-to-hole. If you need to redesign it (and can get away with it), don't put the holes within 1-diameter of the edge. I.e. make the holes smaller, or bring them in closer away from the edge.

 

[saplanti]

I guess what is attached to the system is important in the behaviour of the bracket. You had better include it into your system analysis. If the connected system has its own motion sideways that may take the bracet connections into bending. The bolted area is the weakest and failure is expected there under bending of bracket plate at the connections.
Your case may or may not suit this, however this tells you that you can not eliminate the connected mass (rigid or flexible) to solve the dynamic system.

Additionally, roughly you need to be away from natural frequencies around 10-25% depending on the accuracy of your analysis to prevent resonance.

Hope it helps.

Ibrahim Demir